Visual image-based aero-engine vibration deformation characteristic measurement system and method

Abstract

The invention relates to the technical field of aero-engine testing, in particular to a visual image-based aero-engine vibration deformation characteristic measurement system and method. The measuring system comprises a multi-view-angle high-speed camera used for acquiring image data of an engine test piece at multiple view angles; the binocular vision measurement system is used for acquiring image data of a measurement area of the engine test piece; the acceleration sensors are used for measuring vibration level data of the engine in real time; the camera controller triggers and controls the binocular vision measurement system and the multi-view high-speed camera to carry out image acquisition; and the analysis system is used for carrying out vibration modal quantitative analysis based on image data of the binocular vision measurement system and carrying out modal shape qualitative analysis based on high-speed camera image data of different visual angles. According to the method, the vibration deformation state of the engine can be displayed in quantitative and qualitative modes at the same time, and powerful support is provided for performance evaluation of the aero-engine.

Description

Technical Field

[0001] This invention relates to the field of aero-engine testing technology, and more specifically, to a system and method for measuring the vibration deformation characteristics of aero-engines based on visual images. Background Technology

[0002] Due to the precision and complexity of their structure, the high temperatures of their operating environment, and the high-speed rotation, aero-engines require precise measurement of their vibration and deformation characteristics during operation, especially during the research and development phase. Vibration deformation refers to the cyclic response and deformation of a structure under external excitation or its own operating loads. Vibration deformation characteristic data is crucial for evaluating the rationality of the overall engine structural design and strength design, as well as the stability of bench constraints.

[0003] The measurement of vibration deformation of aero-engines usually adopts the static modal testing method. This method involves selecting some measuring points to install vibration sensors when the engine is stationary, applying excitation by hammering or vibrating agitator, and then using a dynamic data acquisition system to record the data of the vibration measuring points and calculate the engine mode shape accordingly.

[0004] The advantage of static modal testing is that it can obtain data with relatively small errors. However, its disadvantage is that it often requires the installation of a large number of vibration measuring points to calculate the engine's modal shapes. In addition, the engine's modal parameters are easily affected by temperature and pressure, resulting in certain differences between the engine's operating state and its static modes. Therefore, the modal shape results obtained under static conditions cannot be directly used to replace the modal shapes under operating conditions.

[0005] Digital image measurement technology is a non-contact, full-field measurement method that can acquire engine vibration deformation without installing a large number of vibration sensors. This technology simplifies the test system setup process and significantly improves measurement efficiency.

[0006] However, when the entire aero-engine operates under test conditions, the displacement of its external brake is often less than 1 millimeter, while the image displacement is even less than 1 pixel. This tiny displacement, which is imperceptible to the naked eye, poses a great challenge to determining the vibration and deformation of the engine using image measurement technology. Summary of the Invention

[0007] The purpose of this invention is to provide a system and method for measuring the vibration deformation characteristics of aero-engines based on visual images, thereby solving the problem that existing technologies are unable to measure and obtain the engine modal vibration modes under operating conditions.

[0008] To achieve the above objectives, the present invention provides a vision-based system for measuring the vibration deformation characteristics of aero-engines, comprising a multi-view high-speed camera, a binocular vision measurement system, several accelerometers, a camera controller, and an analysis system.

[0009] The multi-view high-speed camera includes high-speed cameras with multiple views, used to acquire image data of the engine test piece from multiple views and send it to the analysis system;

[0010] The binocular vision measurement system is used to acquire image data of the measurement area of ​​the engine test piece and send it to the analysis system;

[0011] The aforementioned acceleration sensors are arranged on the engine test specimen to measure the engine's vibration level data in real time and send it to the analysis system.

[0012] The camera controller, based on the camera trigger signal of the analysis system, simultaneously triggers and controls the binocular vision measurement system and the multi-view high-speed camera to acquire images.

[0013] The analysis system analyzes the vibration level data and determines whether the shooting trigger conditions are met. If they are met, it sends a camera trigger signal to the camera controller. It performs quantitative analysis of vibration modes based on image data from the binocular vision measurement system and qualitative analysis of mode shapes based on high-speed camera image data from different perspectives.

[0014] In some embodiments, the analysis system further includes a real-time analysis system and an image analysis system:

[0015] The real-time analysis system is connected to the acceleration sensor and the camera controller respectively, acquires and analyzes the vibration level data of the engine measured in real time by the acceleration sensor, and determines whether to send a camera trigger signal to the camera controller.

[0016] The image analysis system is connected to a multi-view high-speed camera and a binocular vision measurement system. It performs quantitative analysis of vibration modes based on the image data acquired from the binocular vision measurement system, and qualitative analysis of mode shapes based on the image data acquired from the high-speed camera at different viewpoints.

[0017] In some embodiments, the plurality of acceleration sensors are arranged at the fulcrum of the engine test specimen.

[0018] In some embodiments, when the engine test piece is a single-rotor system, the acceleration sensor is installed at the horizontal and vertical positions of the front and rear supports of the rotor.

[0019] In some embodiments, when the engine test piece is a multi-rotor system, the acceleration sensor is installed at the front support point of each rotor, arranged in both the horizontal and vertical directions of the rotor system.

[0020] In some embodiments, the binocular vision measurement system further includes a plurality of optical targets distributed along the contour of the engine test specimen.

[0021] In some embodiments, the shooting triggering conditions include: the engine reaching a preset speed, the engine entering a surge mode, and a specific area of ​​the engine reaching a specific vibration intensity.

[0022] In some embodiments, the shooting triggering conditions include:

[0023] The engine speed is within a preset range of the engine's critical speed;

[0024] The critical speed of the engine is determined by the fundamental frequency vibration level corresponding to the acceleration sensor during the engine's acceleration and deceleration process;

[0025] The acceleration sensors are horizontal and vertical acceleration sensors located at the front support point of the low-pressure rotor of the engine.

[0026] In some embodiments, the shooting triggering conditions include:

[0027] Vibration level data from any of the accelerometers exceeds a preset threshold.

[0028] In some embodiments, the shooting triggering condition further includes:

[0029] The engine test piece is in a stable operating speed state or the engine test piece is in a speed change stage with a rate of change lower than the preset rate.

[0030] In some embodiments, the binocular vision measurement system simultaneously captures images of the measurement area of ​​the engine test piece using a pair of cameras and sends the obtained image data to an analysis system.

[0031] The analysis system calibrates the camera of the binocular vision measurement system, extracts and analyzes the response features of the obtained image data from the binocular vision measurement system, and obtains the vibration deformation magnitude of the engine test piece.

[0032] In some embodiments, the response features extracted by the analysis system are the spatial positions of the optical targets in each pair of photographs:

[0033] By continuously shooting and extracting data, the spatial coordinate response characteristics of the corresponding target are obtained over time, thereby obtaining the spatial time history response of the target distributed on the outline of the test piece, and quantitatively analyzing the degree of vibration deformation of the engine test piece.

[0034] In some embodiments, the multi-view high-speed camera includes a high-speed camera for capturing the horizontal vibration view of the test specimen, a high-speed camera for capturing the vertical vibration view of the test specimen, and a high-speed camera for capturing the lateral vibration view of the test specimen.

[0035] All the multi-view high-speed cameras are synchronously triggered by the camera controller, and the sampling frame rate remains consistent.

[0036] In some embodiments, the analysis system performs multi-scale image processing, single-frequency vibration extraction, and video reconstruction on high-speed camera image data from different perspectives to conduct qualitative analysis of the modal shapes of the engine test component.

[0037] The multi-scale image processing involves filtering the original video using an operable complex pyramid filter to obtain images of different scales and orientations.

[0038] The single-frequency vibration extraction is performed by extracting single-frequency vibrations from the decomposed components of different image phase differences.

[0039] The video reconstruction involves amplifying the single-frequency phase difference signal and performing image processing in the frequency and spatial domains, then sorting the processed images according to time to form a video.

[0040] In some embodiments, the analysis system performs multi-scale image processing on high-speed camera image data from different perspectives, further including:

[0041] Perform a two-dimensional fast Fourier transform on each frame of the video to convert it to the frequency domain;

[0042] By using the basis functions of the operable complex pyramid, the frequency domain image is filtered to generate multiple sets of frequency domain images of different scales, angles, and frequency bands;

[0043] The filtered frequency domain image is subjected to inverse Fourier transform to obtain a complex image matrix, which corresponds to different sizes, frequency domain directions and angles;

[0044] Calculate the phase angle for each element of the complex image matrix to obtain the phase angle matrix;

[0045] Using the phase of the first image as a reference, calculate the phase difference of other images relative to the reference image.

[0046] In some embodiments, the analysis system performs single-frequency vibration extraction on high-speed camera image data from different perspectives, further including:

[0047] If no near-frequency mode exists, a bidirectional bandpass filter is used to obtain the phase difference signal of the single-frequency vibration signal;

[0048] If near-frequency modes exist, the phase difference signal of the single-frequency vibration signal is obtained by using a system feature extraction algorithm.

[0049] In some embodiments, the analysis system performs video reconstruction on high-speed camera image data from different perspectives, further including:

[0050] Multiply the filtered phase difference signal by the amplification factor to obtain the amplified phase difference;

[0051] The phase angle is amplified and its natural base is used as a power function. The processed phase angle is then subjected to a fast Fourier transform in combination with the original amplitude to obtain the frequency domain image.

[0052] All frequency domain images are superimposed, and a fast inverse Fourier transform is performed on the superimposed frequency domain image to obtain the processed image.

[0053] The processed images are arranged in a time sequence to form a video.

[0054] To achieve the above objectives, the present invention provides a method for measuring the vibration deformation characteristics of aero-engines based on visual images, implemented using the aforementioned visual image-based aero-engine vibration deformation characteristic measurement system, and comprising the following steps:

[0055] Several acceleration sensors were placed on the engine test specimen;

[0056] The analysis system determines the shooting trigger conditions;

[0057] Engine test piece operation;

[0058] The analysis system analyzes the vibration level data measured in real time by the accelerometer and determines whether the shooting trigger conditions are met. If the conditions are met, it sends a camera trigger signal to the camera controller. The camera controller simultaneously triggers and controls the binocular vision measurement system and the multi-view high-speed camera to acquire images.

[0059] A binocular vision measurement system acquires image data of the measurement area of ​​the engine test piece, and a multi-view high-speed camera acquires image data of the engine test piece from multiple perspectives;

[0060] The analysis system performs quantitative vibration mode analysis based on image data from a binocular vision measurement system and qualitative mode analysis based on image data from high-speed cameras at different viewpoints.

[0061] This invention provides a vision-based image-based system and method for measuring the vibration and deformation characteristics of aero-engines. By constructing an image measurement system for aero-engines in operation, the system enables the measurement of vibration and deformation during engine operation. Furthermore, through an engine vibration and deformation visualization method, it is possible to amplify minute vibration and deformation in the image, making its operational modes visible. This not only allows for the quantitative collection and analysis of vibration and deformation data of the engine stator casing structure in operation, but also provides a clear qualitative representation of the engine's vibration and deformation state, offering strong support for aero-engine performance evaluation. Attached Figure Description

[0062] The above and other features, properties and advantages of the present invention will become more apparent from the following description taken in conjunction with the accompanying drawings and embodiments, in which the same reference numerals always denote the same features, wherein:

[0063] Figure 1 A schematic diagram of a vision-based image-based aero-engine vibration deformation characteristic measurement system according to an embodiment of the present invention is disclosed.

[0064] Figure 2 A schematic diagram of a vision-based image-based aero-engine vibration deformation characteristic measurement system according to an embodiment of the present invention is disclosed.

[0065] Figure 3 A flowchart of a vision-based image-based aero-engine vibration deformation feature measurement system according to an embodiment of the present invention is disclosed.

[0066] Figure 4 A schematic flowchart illustrating the image data analysis and processing of a multi-view high-speed camera according to an embodiment of the present invention is disclosed.

[0067] Figure 5 A schematic diagram of the basis functions of an operable complex pyramid according to an embodiment of the present invention is disclosed.

[0068] The meanings of the labels in the figures are as follows:

[0069] 10-view high-speed camera;

[0070] 11. First high-speed camera;

[0071] 12 Second high-speed camera;

[0072] 13. Third high-speed camera;

[0073] 20 Binocular vision measurement system;

[0074] 30 accelerometers;

[0075] 40 camera controllers;

[0076] 50 Analysis System;

[0077] 51 Real-time Analysis System;

[0078] 52 Image Analysis System;

[0079] 60 engine test piece. Detailed Implementation

[0080] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention.

[0081] The key technical challenge in obtaining vibration and deformation characteristics of the engine under full operating conditions lies in:

[0082] First, high-speed cameras often have limited storage capacity, so it is necessary to capture the corresponding image data in a timely manner at a specific speed (or moment) when the engine undergoes overall deformation. Therefore, it is necessary to establish the relationship between the high-speed camera shooting time and the engine operating status.

[0083] Secondly, the displacement of the entire aircraft engine during operation is very small, and its image displacement is often less than 1 pixel. It is difficult to analyze engine vibration and deformation using such a small displacement.

[0084] Finally, engine vibration deformation is complex and may contain near-frequency or dense modes, making it very difficult to distinguish between near-frequency or dense modes in vibration images.

[0085] Among them, image measurement refers to the process of identifying the spatial location and deformation characteristics of specific markings on the surface of a structure using digital image technology; operating state describes the state of a mechanical device or equipment operating according to a set rule, such as the operating state of an aircraft engine; and vibration mode refers to the vibration deformation state of a structure under the excitation of a specific natural frequency.

[0086] This invention proposes a visual image-based system and method for measuring the vibration deformation characteristics of aero-engines. By building an image measurement system for aero-engines in operation, it enables accurate measurement of vibration deformation during engine operation. Furthermore, through a visualization technology for engine vibration deformation, it can amplify minute vibration deformations in the image, thereby achieving a visual representation of the engine's operating modes.

[0087] Figure 1 A schematic diagram of a vision-based image-based aero-engine vibration deformation feature measurement system according to an embodiment of the present invention is disclosed. Figure 2A schematic diagram of a vision-based image-based aero-engine vibration deformation feature measurement system according to an embodiment of the present invention is shown, such as... Figure 1 and Figure 2 As shown, the present invention proposes a vision-based image-based system for measuring the vibration deformation characteristics of aero-engines, comprising a multi-view high-speed camera 10, a binocular vision measurement system 20, several acceleration sensors 30, a camera controller 40, and an analysis system 50.

[0088] The multi-view high-speed camera 10 includes high-speed cameras with multiple views, used to acquire image data of the engine test piece 60 from multiple views and send it to the analysis system 50.

[0089] The binocular vision measurement system 20 is used to acquire image data of the measurement area of ​​the engine test piece 60 and send it to the analysis system 50.

[0090] The plurality of acceleration sensors 30 are arranged on the engine test piece 60 for measuring the vibration level data of the engine in real time and sending it to the analysis system 50.

[0091] The camera controller 40, based on the camera trigger signal of the analysis system 50, simultaneously triggers and controls the binocular vision measurement system 20 and the multi-view high-speed camera 10 to acquire images.

[0092] The analysis system 50 analyzes the vibration level data and determines whether the shooting trigger conditions are met. If they are met, it sends a camera trigger signal to the camera controller 40. It performs quantitative analysis of vibration modes based on the image data from the binocular vision measurement system 20 and qualitative analysis of mode shapes based on high-speed camera image data from different perspectives.

[0093] exist Figure 1 and Figure 2 In the illustrated embodiment, the analysis system 50 further includes a real-time analysis system 51 and an image analysis system 52:

[0094] The real-time analysis system 51 is connected to the acceleration sensor 30 and the camera controller 40 respectively, acquires and analyzes the vibration level data of the engine measured in real time by the acceleration sensor, and determines whether to send a camera trigger signal to the camera controller.

[0095] The image analysis system 52 is connected to the multi-view high-speed camera 10 and the binocular vision measurement system 20, respectively. It performs quantitative analysis of vibration modes based on the image data acquired from the binocular vision measurement system 20, and qualitative analysis of mode shapes based on the high-speed camera image data acquired from different viewpoints.

[0096] Typically, the real-time analysis system 51 and the image analysis system 52 are mainly composed of computers.

[0097] exist Figure 1 and Figure 2 In the illustrated embodiment, the multi-view high-speed camera 10 includes a first high-speed camera 11 for capturing the horizontal vibration view of the test specimen, a second high-speed camera 12 for capturing the vertical vibration view of the test specimen, and a third high-speed camera 13 for capturing the lateral vibration view of the test specimen.

[0098] A high-speed camera is a camera capable of continuously shooting and recording digital images. In this embodiment, a high-speed camera is defined as a camera with a shooting frame rate of more than 500 FPS (Frames Per Second), which ensures that the camera can capture subtle changes in a high-speed dynamic process.

[0099] Furthermore, to ensure consistency in the analysis, the three high-speed cameras are externally synchronized via camera controller 40 to maintain a consistent sampling frame rate.

[0100] Furthermore, the multi-view high-speed camera 10 also includes a matching lighting device to optimize the shooting environment and ensure that clear and accurate image data can be obtained from different perspectives.

[0101] Figure 3 A flowchart of a vision-based image-based aero-engine vibration deformation feature measurement system according to an embodiment of the present invention is disclosed below. Figures 1 to 3 The workflow of the vision image-based aero-engine vibration deformation feature measurement system proposed in this invention will be further elaborated.

[0102] The specific analysis process of the vision-based aero-engine vibration deformation characteristic measurement system proposed in this invention includes: measurement point layout, determination of shooting trigger conditions, acquisition of engine operation image data, quantitative analysis of vibration modes based on binocular vision measurement system, and qualitative analysis of modal shapes based on high-speed camera image processing from different perspectives.

[0103] The following describes the measurement point setup process. The specific steps in this process are to place accelerometers and optical targets on the test specimen to provide monitoring points for subsequent data acquisition and analysis, ensuring the accuracy and reliability of the measurement results.

[0104] The number of acceleration sensors 30 is relatively small, and they are mainly arranged at the support points of the engine rotor test piece 60. Each support point includes monitoring in both horizontal and vertical directions to facilitate real-time monitoring of vibration levels.

[0105] An acceleration sensor 30 is typically placed at the front support of the rotor of an aircraft engine for safety monitoring.

[0106] When the engine test piece is a single rotor system, the acceleration sensor 30 is installed at the horizontal and vertical positions of the front and rear supports of the rotor.

[0107] When the engine test piece is a multi-rotor system, the acceleration sensor 30 is installed at the front support point of each rotor and arranged in both the horizontal and vertical directions of the rotor system.

[0108] The binocular vision measurement system 20 also includes several optical targets:

[0109] The large number of optical targets distributed along the outline of the engine test piece gives the engine strong optical characteristics.

[0110] More specifically, the visible contours of the engine casing can be used to arrange optical feature targets.

[0111] The following describes the process for determining the camera's shooting trigger conditions.

[0112] To establish the relationship between the shooting time of the multi-view high-speed camera and the binocular vision measurement system and the engine operating state, this invention constructs a triggering method based on real-time monitoring and indication of vibration level by an acceleration sensor. This method can measure the vibration level of the engine at different speeds in real time, thereby determining the engine's resonant speed. The multi-view high-speed camera and binocular vision measurement system can then be triggered and images acquired at the appropriate time.

[0113] To optimize storage space usage and capture as many valid images as possible in a single experiment, the constructed multi-view high-speed camera and binocular vision measurement system employs various trigger acquisition strategies. These strategies allow only critical data to be saved, thus avoiding unnecessary data consumption of storage resources.

[0114] The shooting trigger conditions include, but are not limited to: the engine reaching a preset speed, the engine entering surge mode, and a specific area of ​​the engine reaching a specific vibration intensity.

[0115] The commonly used shooting triggering conditions include the following two types:

[0116] The first type is where the engine speed is within a preset range of the engine's critical speed.

[0117] The second scenario is when the vibration level data from any one of the acceleration sensors exceeds a preset threshold.

[0118] More specifically, for the first shooting trigger condition, the vibration of the entire engine is often near the critical speed of the rotor. The critical speed of the engine can be determined by the vibration level of the fundamental frequency (i.e., the response characteristic spectrum corresponding to the first harmonic frequency of the corresponding speed) of the acceleration sensor during the engine's acceleration and deceleration process. When crossing the critical speed, the fundamental frequency of vibration often reaches its peak value. Therefore, setting the trigger condition to be near the critical speed is sufficient to capture images of the vibration and deformation of the entire engine.

[0119] The first shooting trigger condition of the present invention may also include: determining the whole machine resonance speed according to the engine lifting and turning process, and setting the trigger condition within a preset range based on the resonance speed.

[0120] For the first shooting trigger condition, the acceleration sensor 30 is located in the horizontal and vertical directions at the front support point (or bearing seat number one) of the engine's low-pressure rotor.

[0121] More specifically, for the second shooting trigger condition, since there are many factors that cause the engine vibration level to exceed the limit, including surge, local modal resonance, etc., the shooting trigger condition can be set by using the vibration level data fed back by the acceleration sensor 30.

[0122] For any of the monitored accelerometers 30, recording is triggered when the accelerometer exceeds the threshold.

[0123] For example, the threshold value of the trigger displacement peak displayed by the accelerometer is preset to 0.05mm. If the threshold value of the trigger displacement displayed by the accelerometer exceeds 0.05mm, the shooting will be triggered.

[0124] For example, the second shooting triggering condition of the present invention may also include: a pre-set vibration margin level, such as a threshold of 0.1 mm for the fundamental frequency amplitude, and triggering shooting when the fundamental frequency amplitude exceeds 0.1 mm.

[0125] To further optimize the analysis results, the engine should be operating at a stable speed or with slow speed changes when data acquisition is triggered. Therefore, the first or second shooting trigger condition, or other shooting trigger conditions, must also meet the following additional shooting trigger conditions:

[0126] The engine test piece is in a stable operating speed state or the engine test piece is in a speed change stage below the preset change rate.

[0127] The following describes the engine operation process and the process of acquiring engine operation image data.

[0128] The engine test piece 60 is tested according to the predetermined test pattern. The acceleration sensor 30 monitors and feeds back the vibration level data of key positions of the engine in real time. The camera controller 40 sends a trigger signal to the multi-view high-speed camera 10 and the binocular vision measurement system 20 according to the set trigger conditions. Subsequently, the multi-view high-speed camera 10 and the binocular vision measurement system 20 are responsible for recording the corresponding image / video data.

[0129] The following describes the quantitative analysis process of vibration modes based on a binocular vision measurement system.

[0130] The binocular vision measurement system 20 uses a pair of cameras to simultaneously capture images of the measurement area of ​​the engine test piece, which is used to quantify the engine vibration displacement, and sends the obtained image data to the image analysis system 52.

[0131] The image analysis system 52 calibrates the camera of the binocular vision measurement system 20, extracts and analyzes the response features of the obtained image data from the binocular vision measurement system 20, and obtains the vibration deformation magnitude of the engine test piece 60.

[0132] The response features extracted by the image analysis system 52 are the spatial positions of the optical targets in each pair of photographs:

[0133] By continuously shooting and extracting response feature data, the spatial coordinates of the corresponding target change over time (i.e., spatial time history response) are obtained, thereby obtaining the spatial time history response of the target distributed on the outline of the test piece. This allows for precise quantification of the vibration deformation magnitude of the engine test piece 60 and quantitative analysis of the degree of vibration deformation of the engine test piece 60.

[0134] In this embodiment, DIC (Digital Image Correlation) software is used to analyze the target in the image in order to extract the three-dimensional spatial coordinates of the deployed target.

[0135] DIC (Digital Image Correlation) is a technique that uses digital images to perform motion or deformation analysis. Currently available software includes VIC-3D and TEMA, and open-source software such as DICe can all meet the requirements described in this invention.

[0136] Before using DIC software to extract 3D spatial coordinates, the intrinsic and extrinsic parameters of the two cameras must be determined through camera calibration.

[0137] The following section will describe the modal shape qualitative analysis process based on high-speed camera image data processing from different perspectives.

[0138] Figure 4A schematic flowchart illustrating the image data analysis and processing of a multi-view high-speed camera according to an embodiment of the present invention is disclosed, such as... Figure 4 As shown, the image analysis system 52 performs multi-scale image processing, single-frequency vibration extraction, and video reconstruction on high-speed camera image data from different perspectives, and performs qualitative analysis on the modal shapes of the engine test piece:

[0139] The multi-scale image processing includes using an operable complex pyramid filtering technique to filter the original video and obtain images of different scales and orientations;

[0140] The single-frequency vibration extraction includes performing single-frequency vibration extraction on the decomposed components of different image phase differences;

[0141] The video reconstruction includes amplifying the single-frequency phase difference signal and superimposing it in the frequency domain, then transforming it back to the spatial domain to complete image processing, and finally sorting the processed images according to time to generate a video.

[0142] Through the above processing, the acquired video data can display the vibration and deformation videos of the engine test component from different perspectives.

[0143] Furthermore, by analyzing the frequency domain characteristics of the real-time vibration response of the engine test piece, specific vibration response features can be extracted. For example, if the fundamental frequency of the low-pressure rotor vibration is 50Hz and it maintains a constant speed, then the frequency band of interest is [45Hz 55Hz].

[0144] The following is a detailed explanation of the image data analysis process, covering the steps of multi-scale image processing, single-frequency vibration extraction, and video reconstruction.

[0145] This invention introduces an image motion magnification method and applies it to process engine operation videos, thereby amplifying minute vibrations and visualizing the overall engine mode shape.

[0146] The image analysis system 52 performs multi-scale image processing on high-speed camera image data from different perspectives, further including:

[0147] Perform a two-dimensional fast Fourier transform on each frame of the video to convert it to the frequency domain;

[0148] By using the basis functions of the operable complex pyramid, the frequency domain image is filtered to generate multiple sets of frequency domain images of different scales, angles, and frequency bands;

[0149] The filtered frequency domain image is subjected to inverse Fourier transform to obtain a complex image matrix, which corresponds to different sizes, frequency domain directions and angles;

[0150] Calculate the phase angle for each element of the complex image matrix to obtain the phase angle matrix;

[0151] Using the phase of the first image as a reference, calculate the phase difference of other images relative to the reference image.

[0152] More specifically, each frame of the video is first subjected to a two-dimensional fast Fourier transform to transform the image to the spatial frequency domain.

[0153] Figure 5 A schematic diagram of the basis functions of an operable complex pyramid according to an embodiment of the present invention is disclosed, such as... Figure 5 As shown, based on the image size, the basis functions of the operable complex pyramid are used to spatially filter the frequency domain image to obtain multiple sets of frequency domain images at different scales, angles, and frequency bands from high frequency to low frequency.

[0154] The basis functions of the operable complex pyramid, as a composite function combining complex sine functions and Gaussian windows, simultaneously achieve locality analysis in both the frequency and spatial domains. Specific methods are detailed in the reference "SIMONCELLI, EP, AND FREEMAN, WT 1995. The steerable pyramid: a flexible architecture for multi-scale derivative computation. In Proceedings of the 1995 International Conference on Image Processing (Vol. 3) - Volume 3 - Volume 3, IEEE Computer Society, Washington, DC, USA, ICIP '95, 3444–."

[0155] This invention innovatively introduces image processing technology based on directional operable pyramids to perform motion decomposition and video reconstruction, thereby realizing the visualization of vibration and deformation during engine operation.

[0156] Next, an inverse Fourier transform is performed on the filtered frequency domain image to obtain complex image matrices, which correspond to different sizes, frequency domain directions, and angles.

[0157] The image matrix of the i-th image after multi-scale image processing and filtering at a certain size and angle can be represented as:

[0158]

[0159] in, Z represents the sub-image matrix corresponding to a certain size and angle after multi-scale filtering of the i-th image (e.g., a sub-image with scale level 1 and angle of 0 degrees).ij Z represents the complex grayscale values ​​in the transformed matrix. The matrix has m rows, therefore the first column vector is Z. 11 Z 21 ,…,Z m1 The matrix has n columns, therefore the first row is the vector Z. 11 Z 12 ,…,Z 1n .

[0160] For each complex element in the image matrix, calculate the phase angle to obtain the phase angle matrix. It can be represented as:

[0161]

[0162] Phase angle matrix Each element is between -π and π, and the phase change over time represents local motion in the image.

[0163] The complex image phase of the first image For reference, the change in phase is represented by the phase difference, that is, the phase difference Δ(t) between all images in the engine video and the reference image is calculated. i The corresponding expression is:

[0164]

[0165] in, It is the phase angle matrix of the i-th image, t i It is the trigger time of the i-th image.

[0166] The extraction and processing of single-frequency signals needs to be discussed based on whether near-frequency modes exist, and different methods should be used for different situations.

[0167] This invention proposes a dense mode decomposition method to separate near-frequency vibration signals and extract single-mode response features from complex vibration videos.

[0168] The image analysis system 52 performs single-frequency vibration extraction on high-speed camera image data from different perspectives, further including:

[0169] If no near-frequency mode exists, a bidirectional bandpass filter is used to obtain the phase difference signal of the single-frequency vibration signal;

[0170] If near-frequency modes exist, the phase difference signal of the single-frequency vibration signal is obtained by using a system feature extraction algorithm.

[0171] More specifically, for cases where near-frequency modes are absent, a bidirectional bandpass filter is used for amplification, avoiding frequency domain phase distortion caused by the filter and achieving separation of single-frequency vibration deformation. The specific steps are as follows:

[0172] First, determine the engine's vibration frequency, i.e., the frequency range of interest, and then analyze the phase difference signal Δ(t). i Bidirectional time-domain filtering is performed, with the passband of the bandpass filter set to the frequency band of interest. The bandwidth of the non-loss-of-generality filter can be set to 10Hz. The signal is filtered twice, in both the forward and reverse directions, to obtain the filtered phase difference signal.

[0173] For cases involving near-frequency modes, a system feature extraction algorithm is used for separation, achieving the separation of near-frequency signals and the extraction of single-mode signals. The specific steps are as follows:

[0174] Construct the sequence matrix T(0) of the phase difference function.

[0175] T(0)=[Δ(1) Δ(2) … Δ(l)] (4)

[0176] Singular value decomposition of the sequence matrix yields:

[0177] T(0) = USV T (5)

[0178] Where U is the matrix composed of the left singular vectors in the singular value decomposition result, and satisfies U T U = I, where I is an identity matrix with the same dimension and columns as U; V is a matrix composed of right singular vectors, and satisfies V T V = I, where I is an identity matrix with the same dimension and columns as V; S is a singular value matrix, a square matrix with positive values ​​only in its diagonal elements.

[0179] By truncating the obtained singular value S, we can obtain:

[0180]

[0181] In the above formula, The singular values ​​are the first two non-zero singular values ​​in the original singular value matrix S.

[0182] Construct the system matrix A and the observation matrix C:

[0183]

[0184]

[0185] In the above formula, The singular vectors corresponding to the first two non-zero singular values;

[0186] E = [I0], where I is the identity matrix and 0 is the zero matrix.

[0187] A single-frequency time-domain sequence can be calculated using equation (9):

[0188]

[0189] In the above formula This is the phase difference signal after being filtered by a single frequency component.

[0190] Video reconstruction

[0191] The analysis system performs video reconstruction on high-speed camera image data from different perspectives, further including:

[0192] Multiply the filtered phase difference signal by the amplification factor to obtain the amplified phase difference;

[0193] The phase angle is amplified and its natural base is used as a power function. The processed phase angle is then subjected to a fast Fourier transform in combination with the original amplitude to obtain the frequency domain image.

[0194] All frequency domain images are superimposed, and a fast inverse Fourier transform is performed on the superimposed frequency domain images to obtain the processed image.

[0195] The processed images are arranged in a time sequence to form a video.

[0196] More specifically, for the extracted phase difference signal To amplify the signal, multiply the filtered phase difference function by the amplification factor α, as needed:

[0197]

[0198] Where α is the amplification factor, which is usually taken as 50, 100, etc.

[0199] The amplified phase angle, taken as a power function of the natural base, combined with the original amplitude, corresponds to the following expression:

[0200]

[0201] In the above formula, Ψ(t) i () represents the amplitude function corresponding to the phase function, where upright i is the imaginary unit and italic i is the variable for image counting. Let be the complex image matrix of the i-th image.

[0202] right By performing a Fast Fourier Transform, a frequency domain image corresponding to the spatial size and angle can be obtained.

[0203] The processed image can be obtained by superimposing these frequency domain images and performing a fast inverse Fourier transform.

[0204] The processed image results are arranged in time sequence and formed into a video, thus completing the video reconstruction.

[0205] The processed vibration deformation video images from three different perspectives reveal the modal shapes of the test specimen from various viewpoints, qualitatively demonstrating its vibration deformation state. The deformation response characteristics of the test specimen, acquired through a binocular vision measurement system, quantitatively describe its vibration deformation state. Combining qualitative and quantitative deformation analysis allows for accurate identification of the overall deformation characteristics of the engine.

[0206] Based on the aforementioned vision-based image-based aero-engine vibration deformation feature measurement system, this invention also proposes a vision-based image-based aero-engine vibration deformation feature measurement method, comprising the following steps:

[0207] Several acceleration sensors were placed on the engine test specimen;

[0208] The analysis system determines the shooting trigger conditions;

[0209] Engine test piece operation;

[0210] The analysis system analyzes the vibration level data measured in real time by the accelerometer and determines whether the shooting trigger conditions are met. If the conditions are met, it sends a camera trigger signal to the camera controller. The camera controller simultaneously triggers and controls the binocular vision measurement system and the multi-view high-speed camera to acquire images.

[0211] A binocular vision measurement system acquires image data of the measurement area of ​​the engine test piece, and a multi-view high-speed camera acquires image data of the engine test piece from multiple perspectives;

[0212] The analysis system performs quantitative vibration mode analysis based on image data from a binocular vision measurement system and qualitative mode analysis based on image data from high-speed cameras at different viewpoints.

[0213] The specific implementation details of the vision image-based aero-engine vibration deformation feature measurement method correspond to the workflow of the aforementioned vision image-based aero-engine vibration deformation feature measurement system, so the specific details will not be repeated here.

[0214] The present invention provides a system and method for measuring the vibration deformation characteristics of aero-engines based on visual images, which has the following advantages:

[0215] 1) It can extract the vibration modes under engine operating conditions and visualize minute vibrations using image signal amplification technology;

[0216] 2) It can determine the engine resonance speed, thereby triggering the start-up image acquisition and storage, and saving the engine operation image and video clips of interest;

[0217] 3) It can extract engine mode shapes under stable engine operation conditions, which can be directly used for test runs during engine development to analyze engine vibration and deformation characteristics, provide feedback for engine strength design, and correct finite element models;

[0218] 4) The proposed engine micro-vibration amplification and visualization technology can amplify the tiny image displacement of the engine to the level visible to the naked eye, making it easy to intuitively display and analyze the overall vibration state of the engine during operation.

[0219] Although the methods described above are illustrated and depicted as a series of actions for the sake of simplicity, it should be understood and appreciated that these methods are not limited by the order of the actions, as some actions may occur in a different order and / or concurrently with other actions from the illustrations and descriptions herein or not illustrated and described herein but which may be understood by those skilled in the art, according to one or more embodiments.

[0220] As indicated in this application and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not specifically singular and may include plural forms. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0221] The above embodiments are provided for those skilled in the art to implement or use the present invention. Those skilled in the art can make various modifications or changes to the above embodiments without departing from the inventive concept of the present invention. Therefore, the protection scope of the present invention is not limited to the above embodiments, but should be the maximum scope that conforms to the innovative features mentioned in the claims.

Claims

1. A vision-based image-based system for measuring the vibration and deformation characteristics of an aero-engine, characterized in that, Includes a multi-view high-speed camera, a binocular vision measurement system, several accelerometers, a camera controller, and an analysis system. The multi-view high-speed camera includes high-speed cameras with multiple views, used to acquire image data of the engine test piece from multiple views and send it to the analysis system; The binocular vision measurement system is used to acquire image data of the measurement area of ​​the engine test piece and send it to the analysis system; The aforementioned acceleration sensors are arranged on the engine test specimen to measure the engine's vibration level data in real time and send it to the analysis system. The camera controller, based on the camera trigger signal of the analysis system, simultaneously triggers and controls the binocular vision measurement system and the multi-view high-speed camera to acquire images. The analysis system analyzes the vibration level data and determines whether the shooting trigger conditions are met. If they are met, it sends a camera trigger signal to the camera controller. It performs quantitative analysis of vibration modes based on image data from the binocular vision measurement system and qualitative analysis of mode shapes based on image data from high-speed cameras at different viewpoints.

2. The vision-based image-based aero-engine vibration deformation characteristic measurement system according to claim 1, characterized in that, The analysis system further includes a real-time analysis system and an image analysis system: The real-time analysis system is connected to the acceleration sensor and the camera controller respectively, acquires and analyzes the vibration level data of the engine measured in real time by the acceleration sensor, and determines whether to send a camera trigger signal to the camera controller. The image analysis system is connected to a multi-view high-speed camera and a binocular vision measurement system, respectively. It performs quantitative analysis of vibration modes based on the image data acquired from the binocular vision measurement system, and qualitative analysis of mode shapes based on the image data acquired from the high-speed camera at different viewpoints.

3. The vision-based image-based aero-engine vibration deformation characteristic measurement system according to claim 1, characterized in that, The aforementioned acceleration sensors are arranged at the fulcrum of the engine test piece.

4. The vision-based image-based aero-engine vibration deformation characteristic measurement system according to claim 3, characterized in that, When the engine test piece is a single-rotor system, the acceleration sensor is installed at the horizontal and vertical positions of the front and rear supports of the rotor.

5. The vision-based image-based aero-engine vibration deformation characteristic measurement system according to claim 3, characterized in that, When the engine test piece is a multi-rotor system, the acceleration sensor is installed at the front support point of each rotor, arranged in both the horizontal and vertical directions of the rotor system.

6. The vision-image-based aero-engine vibration deformation characteristic measurement system according to claim 1, characterized in that, The binocular vision measurement system also includes several optical targets distributed along the outline of the engine test piece.

7. The vision-based image-based aero-engine vibration deformation characteristic measurement system according to claim 1, characterized in that, The shooting trigger conditions include: the engine reaching a preset speed, the engine entering surge mode, and a specific area of ​​the engine reaching a specific vibration intensity.

8. The vision-based image-based aero-engine vibration deformation characteristic measurement system according to claim 1, characterized in that, The shooting trigger conditions include: The engine speed is within a preset range of the engine's critical speed; The critical speed of the engine is determined by the fundamental frequency vibration level corresponding to the acceleration sensor during the engine's acceleration and deceleration process; The acceleration sensors are horizontal and vertical acceleration sensors located at the front support point of the low-pressure rotor of the engine.

9. The vision-image-based aero-engine vibration deformation characteristic measurement system according to claim 1, characterized in that, The shooting trigger conditions include: Vibration level data from any of the accelerometers exceeds a preset threshold.

10. The vision-based image-based aero-engine vibration deformation characteristic measurement system according to any one of claims 1, 7, 8, or 9, characterized in that, The shooting triggering conditions also include: The engine test piece is in a stable operating speed state or the engine test piece is in a speed change stage with a rate of change lower than the preset rate.

11. The vision-based image-based aero-engine vibration deformation characteristic measurement system according to claim 1, characterized in that, The binocular vision measurement system uses a pair of cameras to simultaneously capture images of the measurement area of ​​the engine test piece and sends the obtained image data to the analysis system. The analysis system calibrates the camera of the binocular vision measurement system, extracts and analyzes the response features of the obtained image data from the binocular vision measurement system, and obtains the vibration deformation magnitude of the engine test piece.

12. The vision-based image-based aero-engine vibration deformation characteristic measurement system according to claim 11, characterized in that, The response features extracted by the analysis system are the spatial positions of the optical targets in each pair of photographs: By continuously shooting and extracting data, the spatial coordinate response characteristics of the corresponding target are obtained over time, thereby obtaining the spatial time history response of the target distributed on the outline of the test piece, and quantitatively analyzing the degree of vibration deformation of the engine test piece.

13. The vision-image-based aero-engine vibration deformation characteristic measurement system according to claim 1, characterized in that, The multi-view high-speed camera includes a high-speed camera for capturing the horizontal vibration view of the test specimen, a high-speed camera for capturing the vertical vibration view of the test specimen, and a high-speed camera for capturing the lateral vibration view of the test specimen. All the multi-view high-speed cameras are synchronously triggered by the camera controller, and the sampling frame rate remains consistent.

14. The vision-image-based aero-engine vibration deformation characteristic measurement system according to claim 1, characterized in that, The analysis system performs multi-scale image processing, single-frequency vibration extraction, and video reconstruction on high-speed camera image data from different perspectives, and conducts qualitative analysis of the modal shapes of the engine test specimen. The multi-scale image processing involves filtering the original video using an operable complex pyramid filter to obtain images of different scales and orientations. The single-frequency vibration extraction is performed by extracting single-frequency vibrations from the decomposed components of different image phase differences. The video reconstruction involves amplifying the single-frequency phase difference signal and performing image processing in the frequency and spatial domains, then sorting the processed images according to time to form a video.

15. The vision-based image-based aero-engine vibration deformation characteristic measurement system according to claim 14, characterized in that, The analysis system performs multi-scale image processing on high-speed camera image data from different perspectives, further including: Perform a two-dimensional fast Fourier transform on each frame of the video to convert it to the frequency domain; By using the basis functions of the operable complex pyramid, the frequency domain image is filtered to generate multiple sets of frequency domain images of different scales, angles, and frequency bands; The filtered frequency domain image is subjected to inverse Fourier transform to obtain a complex image matrix, which corresponds to different sizes, frequency domain directions and angles; Calculate the phase angle for each element of the complex image matrix to obtain the phase angle matrix; Using the phase of the first image as a reference, calculate the phase difference of other images relative to the reference image.

16. The vision-image-based aero-engine vibration deformation characteristic measurement system according to claim 14, characterized in that, The analysis system performs single-frequency vibration extraction on high-speed camera image data from different perspectives, further including: If no near-frequency mode exists, a bidirectional bandpass filter is used to obtain the phase difference signal of the single-frequency vibration signal; If near-frequency modes exist, the phase difference signal of the single-frequency vibration signal is obtained by using a system feature extraction algorithm.

17. The vision-based image-based aero-engine vibration deformation characteristic measurement system according to claim 14, characterized in that, The analysis system performs video reconstruction on high-speed camera image data from different perspectives, further including: Multiply the filtered phase difference signal by the amplification factor to obtain the amplified phase difference; The phase angle is amplified and its natural base is used as a power function. The processed phase angle is then subjected to a fast Fourier transform in combination with the original amplitude to obtain the frequency domain image. All frequency domain images are superimposed, and a fast inverse Fourier transform is performed on the superimposed frequency domain image to obtain the processed image. The processed images are arranged in a time sequence to form a video.

18. A method for measuring the vibration deformation characteristics of an aero-engine based on visual images, implemented using the visual image-based aero-engine vibration deformation characteristic measurement system as described in any one of claims 1 to 17, characterized in that, Includes the following steps: Several acceleration sensors were placed on the engine test specimen; The analysis system determines the shooting trigger conditions; Engine test piece operation; The analysis system analyzes the vibration level data measured in real time by the accelerometer and determines whether the shooting trigger conditions are met. If the conditions are met, it sends a camera trigger signal to the camera controller. The camera controller simultaneously triggers and controls the binocular vision measurement system and the multi-view high-speed camera to acquire images. A binocular vision measurement system acquires image data of the measurement area of ​​the engine test piece, and a multi-view high-speed camera acquires image data of the engine test piece from multiple perspectives; The analysis system performs quantitative vibration mode analysis based on image data from a binocular vision measurement system and qualitative mode analysis based on image data from high-speed cameras at different viewpoints.

Images